Hi,

We are trying to simulate particles with Langevin dynamics in a shear flow by Lees-Edwards boundary conditions in MD simulations.

It seems thermal forces and shear flow do not go hand in hand, one example is given here:

https://lammps.sandia.gov/threads/msg28596.html

In Lees-Edwards, the averaged velocity profile must be linear, however, with Langevin we get significant deviations from the linear velocity profile, even after we make a lot of averaging.

Is there a remedy to reconcile Langevin dynamics with Lees-Edwards boundary conditions in a MD simulation?

In advance, I appreciate your kind help!